Australian National University
IPC 2022 featured speaker
Dr Liang Zheng is a Senior Lecturer, CS Futures Fellow and DECRA Fellow in the School of Computing, Australian National University.
He is best known for his contributions in object re-identification, and his recent research interest is data-centered computer vision, where improving leveraging, analysing and improving data instead of algorithms are the of primary concern.
He was/is a co-organiser of the AI City workshop series at CVPR and is an Area Chair for CVPR 2020, ECCV 2020, and ACM Multimedia 2020, 2021, and an Associate Editor for IEEE T-CSVT. He received his B.S. degree (2010) and Ph.D. degree (2015) from Tsinghua University, China.
The Promises of Synthetic Data in Computer Vision
Data collected from the real world is usually viewed as the engine for computer vision research. Their limitations are obvious: expensive, hard to be modified, unknown environmental compositions, etc. In comparison, synthetic data, e.g., from simulation engines, mitigates these drawbacks, but its role is severely underestimated due to the so-called appearance gap.
In this talk, I will answer some interesting questions around some new and critical roles synthetic data takes in the community: what makes a good synthetic training set? How is a synthetic test set validated? When does the synthetic-real domain gap disappear?